C7082-Techniques in Machine learning and AI
Student Id
22302900
Background
The aim of this project is to detect the animals that are living in the Aquariums.we will draw the bounding boxes and detect the classes of the animals. For detection we are going to use the YOLOV5 model.you only look once is the meaning of YOLOV5 is an alogorithm that uses neural networks to provide real time object detection.it will detect between various objects in digital images and videos.And we also use tensorflow to display our graphs.
Objective
Climate change made huge differnce in degradation of Aquatic animals and coral reefs.The preservation of coral reefs and marine life depends on underwater health monitoring. In this project, we'll use computer vision and deep learning to create an aquarium object recognition system.
Methods
Data source
https://public.roboflow.com/object-detection/aquarium
Dataset Details
This dataset consists of 638 photos gathered by Roboflow from two aquariums in the United States.The National Aquarium in Baltimore and the Henry Doorly Zoo in Omaha (both on October 16, 2020). (November 14, 2020). This dataset was collected to identify objects. There are seven classes listed below.
- Fish
- stingray
- jellyfish
- penguin
- shark
- puffin
- starfish
For training this model we are going to use train and split method.That means for train we use 70% of images and for split 20% images and for validation 10% of images.
The reason for using train-split method is used to estimate the performance of the Machine learning alogirthams that are applicable for predication based algorithams and applications.This method is fast and easy procedure to perform such that we can compare our own machine learning model results to machine results.
In [ ]:
#clone YOLOv5
!git clone https://github.com/ultralytics/yolov5 # clone repo
%cd yolov5
%pip install -qr requirements.txt # install dependencies
%pip install -q roboflow
import torch
import os
from IPython.display import Image, clear_output # to display images
print(f"Setup complete. Using torch {torch.__version__} ({torch.cuda.get_device_properties(0).name if torch.cuda.is_available() else 'CPU'})")
Cloning into 'yolov5'...
remote: Enumerating objects: 14995, done.
remote: Total 14995 (delta 0), reused 0 (delta 0), pack-reused 14995
Receiving objects: 100% (14995/14995), 14.02 MiB | 31.54 MiB/s, done.
Resolving deltas: 100% (10286/10286), done.
/content/yolov5
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 184.0/184.0 KB 7.1 MB/s eta 0:00:00
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 62.7/62.7 KB 6.3 MB/s eta 0:00:00
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.6/1.6 MB 48.5 MB/s eta 0:00:00
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 46.5/46.5 KB 3.4 MB/s eta 0:00:00
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 67.8/67.8 KB 8.2 MB/s eta 0:00:00
Preparing metadata (setup.py) ... done
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 54.5/54.5 KB 6.4 MB/s eta 0:00:00
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 138.5/138.5 KB 16.0 MB/s eta 0:00:00
Building wheel for wget (setup.py) ... done
Setup complete. Using torch 1.13.1+cu116 (Tesla T4)
create our database
We must put together a dataset of typical photos with bounding box annotations surrounding the things we wish to detect in order to train our custom model. Additionally, we require a dataset in YOLOv5 format.
For this we are using the roboflow.Roboflow is an open AI that will annotate the images and draw bounding boxes.it makes us easy to train our model.
In [ ]:
from roboflow import Roboflow
rf = Roboflow(model_format="yolov5", notebook="ultralytics")
upload and label your dataset, and get an API KEY here: https://app.roboflow.com/?model=yolov5&ref=ultralytics
In [ ]:
# set up environment to save our dataset
os.environ["DATASET_DIRECTORY"] = "/content/datasets"
In [ ]:
#import the dataset to google colab
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="TypZZgpMSpKK7oNxPV26")
project = rf.workspace("kk-fgzul").project("fish-x35tq")
dataset = project.version(1).download("yolov5")
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/ Requirement already satisfied: roboflow in /usr/local/lib/python3.8/dist-packages (0.2.25) Requirement already satisfied: kiwisolver>=1.3.1 in /usr/local/lib/python3.8/dist-packages (from roboflow) (1.4.4) Requirement already satisfied: urllib3==1.26.6 in /usr/local/lib/python3.8/dist-packages (from roboflow) (1.26.6) Requirement already satisfied: six in /usr/local/lib/python3.8/dist-packages (from roboflow) (1.15.0) Requirement already satisfied: pyparsing==2.4.7 in /usr/local/lib/python3.8/dist-packages (from roboflow) (2.4.7) Requirement already satisfied: PyYAML>=5.3.1 in /usr/local/lib/python3.8/dist-packages (from roboflow) (6.0) Requirement already satisfied: opencv-python-headless>=4.5.1.48 in /usr/local/lib/python3.8/dist-packages (from roboflow) (4.7.0.68) Requirement already satisfied: tqdm>=4.41.0 in /usr/local/lib/python3.8/dist-packages (from roboflow) (4.64.1) Requirement already satisfied: cycler==0.10.0 in /usr/local/lib/python3.8/dist-packages (from roboflow) (0.10.0) Requirement already satisfied: matplotlib in /usr/local/lib/python3.8/dist-packages (from roboflow) (3.2.2) Requirement already satisfied: python-dotenv in /usr/local/lib/python3.8/dist-packages (from roboflow) (0.21.0) Requirement already satisfied: requests in /usr/local/lib/python3.8/dist-packages (from roboflow) (2.25.1) Requirement already satisfied: requests-toolbelt in /usr/local/lib/python3.8/dist-packages (from roboflow) (0.10.1) Requirement already satisfied: idna==2.10 in /usr/local/lib/python3.8/dist-packages (from roboflow) (2.10) Requirement already satisfied: certifi==2022.12.7 in /usr/local/lib/python3.8/dist-packages (from roboflow) (2022.12.7) Requirement already satisfied: Pillow>=7.1.2 in /usr/local/lib/python3.8/dist-packages (from roboflow) (7.1.2) Requirement already satisfied: numpy>=1.18.5 in /usr/local/lib/python3.8/dist-packages (from roboflow) (1.21.6) Requirement already satisfied: python-dateutil in /usr/local/lib/python3.8/dist-packages (from roboflow) (2.8.2) Requirement already satisfied: wget in /usr/local/lib/python3.8/dist-packages (from roboflow) (3.2) Requirement already satisfied: glob2 in /usr/local/lib/python3.8/dist-packages (from roboflow) (0.7) Requirement already satisfied: chardet==4.0.0 in /usr/local/lib/python3.8/dist-packages (from roboflow) (4.0.0) loading Roboflow workspace... loading Roboflow project... Downloading Dataset Version Zip in /content/datasets/fish-1 to yolov5pytorch: 100% [37971512 / 37971512] bytes
Extracting Dataset Version Zip to /content/datasets/fish-1 in yolov5pytorch:: 100%|██████████| 1286/1286 [00:00<00:00, 2274.09it/s]
Develop our model
- img: define the input picture size using img.
- batch: Determine the batch size.
- epochs: specify how many training epochs there are.
- data: The dataset location for our project has been stored.
- Weights: Define a starting path for weights. transmit knowledge gained. In this case, we pick the standard COCO pretrained checkpoint.
- cache: pictures should be cached for quicker training.
In [ ]:
!python train.py --img 416 --batch 16 --epochs 400 --data {dataset.location}/data.yaml --weights yolov5s.pt --cache
train: weights=yolov5s.pt, cfg=, data=/content/datasets/fish-1/data.yaml, hyp=data/hyps/hyp.scratch-low.yaml, epochs=400, batch_size=16, imgsz=416, rect=False, resume=False, nosave=False, noval=False, noautoanchor=False, noplots=False, evolve=None, bucket=, cache=ram, image_weights=False, device=, multi_scale=False, single_cls=False, optimizer=SGD, sync_bn=False, workers=8, project=runs/train, name=exp, exist_ok=False, quad=False, cos_lr=False, label_smoothing=0.0, patience=100, freeze=[0], save_period=-1, seed=0, local_rank=-1, entity=None, upload_dataset=False, bbox_interval=-1, artifact_alias=latest github: up to date with https://github.com/ultralytics/yolov5 ✅ YOLOv5 🚀 v7.0-71-gc442a2e Python-3.8.10 torch-1.13.1+cu116 CUDA:0 (Tesla T4, 15110MiB) hyperparameters: lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=0.05, cls=0.5, cls_pw=1.0, obj=1.0, obj_pw=1.0, iou_t=0.2, anchor_t=4.0, fl_gamma=0.0, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0 ClearML: run 'pip install clearml' to automatically track, visualize and remotely train YOLOv5 🚀 in ClearML Comet: run 'pip install comet_ml' to automatically track and visualize YOLOv5 🚀 runs in Comet TensorBoard: Start with 'tensorboard --logdir runs/train', view at http://localhost:6006/ Downloading https://ultralytics.com/assets/Arial.ttf to /root/.config/Ultralytics/Arial.ttf... 100% 755k/755k [00:00<00:00, 29.8MB/s] Downloading https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s.pt to yolov5s.pt... 100% 14.1M/14.1M [00:00<00:00, 200MB/s] Overriding model.yaml nc=80 with nc=7 from n params module arguments 0 -1 1 3520 models.common.Conv [3, 32, 6, 2, 2] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 18816 models.common.C3 [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 2 115712 models.common.C3 [128, 128, 2] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 3 625152 models.common.C3 [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 1182720 models.common.C3 [512, 512, 1] 9 -1 1 656896 models.common.SPPF [512, 512, 5] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 361984 models.common.C3 [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 90880 models.common.C3 [256, 128, 1, False] 18 -1 1 147712 models.common.Conv [128, 128, 3, 2] 19 [-1, 14] 1 0 models.common.Concat [1] 20 -1 1 296448 models.common.C3 [256, 256, 1, False] 21 -1 1 590336 models.common.Conv [256, 256, 3, 2] 22 [-1, 10] 1 0 models.common.Concat [1] 23 -1 1 1182720 models.common.C3 [512, 512, 1, False] 24 [17, 20, 23] 1 32364 models.yolo.Detect [7, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]] Model summary: 214 layers, 7038508 parameters, 7038508 gradients, 16.0 GFLOPs Transferred 343/349 items from yolov5s.pt AMP: checks passed ✅ optimizer: SGD(lr=0.01) with parameter groups 57 weight(decay=0.0), 60 weight(decay=0.0005), 60 bias albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8)) train: Scanning /content/datasets/fish-1/train/labels... 446 images, 0 backgrounds, 0 corrupt: 100% 446/446 [00:00<00:00, 1920.25it/s] train: New cache created: /content/datasets/fish-1/train/labels.cache train: Caching images (0.2GB ram): 100% 446/446 [00:02<00:00, 149.32it/s] val: Scanning /content/datasets/fish-1/valid/labels... 128 images, 0 backgrounds, 0 corrupt: 100% 128/128 [00:00<00:00, 588.68it/s] val: New cache created: /content/datasets/fish-1/valid/labels.cache val: Caching images (0.1GB ram): 100% 128/128 [00:01<00:00, 91.57it/s] AutoAnchor: 4.66 anchors/target, 0.999 Best Possible Recall (BPR). Current anchors are a good fit to dataset ✅ Plotting labels to runs/train/exp/labels.jpg... Image sizes 416 train, 416 val Using 2 dataloader workers Logging results to runs/train/exp Starting training for 400 epochs... Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 0/399 1.71G 0.1169 0.0445 0.06014 118 416: 100% 28/28 [00:10<00:00, 2.56it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:02<00:00, 1.80it/s] all 128 993 0.00471 0.258 0.00541 0.00131 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 1/399 2.07G 0.09639 0.05816 0.04905 168 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 2.09it/s] all 128 993 0.0147 0.142 0.0158 0.00409 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 2/399 2.07G 0.08398 0.05357 0.04337 189 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:02<00:00, 1.94it/s] all 128 993 0.799 0.104 0.1 0.0292 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 3/399 2.07G 0.0789 0.05171 0.04112 146 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.41it/s] all 128 993 0.152 0.25 0.118 0.0327 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 4/399 2.07G 0.07428 0.05138 0.03638 161 416: 100% 28/28 [00:05<00:00, 5.28it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 2.91it/s] all 128 993 0.683 0.182 0.164 0.0549 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 5/399 2.07G 0.06958 0.04762 0.03148 176 416: 100% 28/28 [00:05<00:00, 4.88it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.12it/s] all 128 993 0.121 0.439 0.18 0.0603 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 6/399 2.07G 0.06616 0.04905 0.02949 168 416: 100% 28/28 [00:07<00:00, 3.84it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 2.31it/s] all 128 993 0.16 0.479 0.225 0.0873 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 7/399 2.07G 0.06266 0.04948 0.02781 158 416: 100% 28/28 [00:05<00:00, 4.81it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.07it/s] all 128 993 0.252 0.482 0.308 0.132 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 8/399 2.07G 0.06072 0.04752 0.02477 122 416: 100% 28/28 [00:05<00:00, 4.75it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.48it/s] all 128 993 0.271 0.492 0.329 0.129 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 9/399 2.07G 0.0604 0.04851 0.02403 181 416: 100% 28/28 [00:05<00:00, 5.00it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.71it/s] all 128 993 0.289 0.483 0.36 0.159 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 10/399 2.07G 0.05844 0.04651 0.02257 102 416: 100% 28/28 [00:07<00:00, 3.80it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.36it/s] all 128 993 0.369 0.481 0.423 0.201 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 11/399 2.07G 0.05714 0.04781 0.02083 129 416: 100% 28/28 [00:05<00:00, 4.83it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.33it/s] all 128 993 0.355 0.528 0.398 0.177 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 12/399 2.07G 0.05633 0.04954 0.02113 140 416: 100% 28/28 [00:06<00:00, 4.43it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 4.00it/s] all 128 993 0.412 0.554 0.466 0.199 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 13/399 2.07G 0.05522 0.04496 0.01974 115 416: 100% 28/28 [00:06<00:00, 4.36it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.61it/s] all 128 993 0.444 0.537 0.465 0.217 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 14/399 2.07G 0.05499 0.04565 0.01846 112 416: 100% 28/28 [00:05<00:00, 4.91it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.69it/s] all 128 993 0.532 0.565 0.522 0.246 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 15/399 2.07G 0.05464 0.0474 0.01701 162 416: 100% 28/28 [00:05<00:00, 4.71it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.44it/s] all 128 993 0.613 0.549 0.592 0.298 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 16/399 2.07G 0.05299 0.04315 0.01769 113 416: 100% 28/28 [00:05<00:00, 5.33it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.651 0.568 0.596 0.255 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 17/399 2.07G 0.05317 0.04575 0.01428 241 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.758 0.598 0.674 0.335 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 18/399 2.07G 0.05128 0.04602 0.01409 115 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.14it/s] all 128 993 0.749 0.598 0.648 0.329 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 19/399 2.07G 0.05223 0.04684 0.01281 130 416: 100% 28/28 [00:06<00:00, 4.57it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.39it/s] all 128 993 0.735 0.576 0.657 0.316 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 20/399 2.07G 0.05192 0.04331 0.01224 175 416: 100% 28/28 [00:07<00:00, 3.79it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.18it/s] all 128 993 0.731 0.612 0.652 0.336 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 21/399 2.07G 0.05058 0.04394 0.01199 123 416: 100% 28/28 [00:05<00:00, 5.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.11it/s] all 128 993 0.728 0.631 0.696 0.35 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 22/399 2.07G 0.05189 0.04381 0.01179 149 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.734 0.607 0.665 0.326 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 23/399 2.07G 0.05108 0.04442 0.01061 86 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.54it/s] all 128 993 0.708 0.671 0.69 0.333 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 24/399 2.07G 0.04938 0.04353 0.009887 175 416: 100% 28/28 [00:05<00:00, 5.30it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.35it/s] all 128 993 0.782 0.592 0.695 0.351 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 25/399 2.07G 0.05064 0.04347 0.009374 214 416: 100% 28/28 [00:06<00:00, 4.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.13it/s] all 128 993 0.765 0.62 0.706 0.331 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 26/399 2.07G 0.0498 0.04446 0.00826 144 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.47it/s] all 128 993 0.705 0.624 0.675 0.326 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 27/399 2.07G 0.04876 0.04218 0.008837 226 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.47it/s] all 128 993 0.743 0.599 0.678 0.344 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 28/399 2.07G 0.04902 0.04333 0.008695 108 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.19it/s] all 128 993 0.737 0.64 0.714 0.348 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 29/399 2.07G 0.0481 0.04336 0.008077 168 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.774 0.664 0.711 0.379 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 30/399 2.07G 0.04933 0.04225 0.007511 133 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.716 0.631 0.691 0.336 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 31/399 2.07G 0.04913 0.04301 0.007104 92 416: 100% 28/28 [00:05<00:00, 5.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.82it/s] all 128 993 0.785 0.637 0.711 0.342 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 32/399 2.07G 0.04846 0.04164 0.007358 175 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.807 0.676 0.752 0.393 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 33/399 2.07G 0.04753 0.04227 0.007231 97 416: 100% 28/28 [00:05<00:00, 5.32it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.45it/s] all 128 993 0.716 0.675 0.705 0.367 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 34/399 2.07G 0.04719 0.04263 0.006953 156 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.778 0.665 0.717 0.377 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 35/399 2.07G 0.04613 0.03955 0.007541 146 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.70it/s] all 128 993 0.767 0.644 0.723 0.359 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 36/399 2.07G 0.04843 0.03919 0.007373 133 416: 100% 28/28 [00:05<00:00, 5.32it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.769 0.621 0.702 0.365 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 37/399 2.07G 0.04829 0.04111 0.006599 217 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.55it/s] all 128 993 0.769 0.669 0.731 0.368 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 38/399 2.07G 0.04774 0.04215 0.006241 210 416: 100% 28/28 [00:05<00:00, 4.88it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.817 0.696 0.772 0.372 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 39/399 2.07G 0.04816 0.04372 0.006233 184 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.94it/s] all 128 993 0.789 0.671 0.758 0.414 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 40/399 2.07G 0.04538 0.03996 0.006558 122 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.823 0.645 0.756 0.404 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 41/399 2.07G 0.04665 0.04128 0.00641 121 416: 100% 28/28 [00:05<00:00, 5.32it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.43it/s] all 128 993 0.773 0.667 0.743 0.381 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 42/399 2.07G 0.04605 0.03931 0.006054 166 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.768 0.686 0.749 0.406 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 43/399 2.07G 0.04601 0.03979 0.006124 111 416: 100% 28/28 [00:05<00:00, 5.33it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.34it/s] all 128 993 0.801 0.675 0.74 0.395 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 44/399 2.07G 0.04611 0.04034 0.006076 60 416: 100% 28/28 [00:05<00:00, 5.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.37it/s] all 128 993 0.799 0.664 0.744 0.397 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 45/399 2.07G 0.04624 0.03954 0.005589 134 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.64it/s] all 128 993 0.751 0.687 0.765 0.404 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 46/399 2.07G 0.04691 0.04189 0.005411 173 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.775 0.686 0.76 0.405 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 47/399 2.07G 0.04476 0.0399 0.005895 173 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.779 0.715 0.776 0.409 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 48/399 2.07G 0.04473 0.04068 0.005242 176 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.59it/s] all 128 993 0.799 0.687 0.763 0.4 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 49/399 2.07G 0.04546 0.04049 0.005776 139 416: 100% 28/28 [00:07<00:00, 3.69it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 2.58it/s] all 128 993 0.812 0.662 0.765 0.395 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 50/399 2.07G 0.04557 0.04078 0.005014 167 416: 100% 28/28 [00:07<00:00, 3.68it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.739 0.694 0.76 0.399 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 51/399 2.07G 0.04471 0.03986 0.005029 113 416: 100% 28/28 [00:05<00:00, 4.96it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.769 0.701 0.761 0.4 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 52/399 2.07G 0.04437 0.03823 0.005286 115 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.786 0.66 0.752 0.411 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 53/399 2.07G 0.04304 0.03887 0.00526 143 416: 100% 28/28 [00:05<00:00, 4.96it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.785 0.673 0.755 0.414 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 54/399 2.07G 0.0439 0.0384 0.004443 182 416: 100% 28/28 [00:09<00:00, 2.96it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.28it/s] all 128 993 0.783 0.679 0.745 0.401 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 55/399 2.07G 0.04329 0.03963 0.004856 122 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.58it/s] all 128 993 0.81 0.682 0.773 0.425 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 56/399 2.07G 0.0425 0.03852 0.004776 141 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.45it/s] all 128 993 0.787 0.714 0.778 0.422 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 57/399 2.07G 0.04389 0.04102 0.004347 224 416: 100% 28/28 [00:07<00:00, 3.86it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.69it/s] all 128 993 0.793 0.707 0.768 0.427 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 58/399 2.07G 0.04349 0.04076 0.004174 134 416: 100% 28/28 [00:05<00:00, 5.01it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.55it/s] all 128 993 0.821 0.669 0.757 0.393 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 59/399 2.07G 0.04428 0.039 0.004639 129 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.768 0.715 0.769 0.419 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 60/399 2.07G 0.04282 0.03838 0.004444 129 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.56it/s] all 128 993 0.82 0.699 0.765 0.418 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 61/399 2.07G 0.04348 0.03922 0.004368 143 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.803 0.662 0.742 0.402 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 62/399 2.07G 0.04288 0.03991 0.00456 257 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.54it/s] all 128 993 0.818 0.636 0.743 0.415 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 63/399 2.07G 0.04309 0.03781 0.004396 170 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.855 0.698 0.771 0.42 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 64/399 2.07G 0.04318 0.04128 0.004684 116 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.785 0.677 0.738 0.394 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 65/399 2.07G 0.04226 0.03831 0.004544 179 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.806 0.686 0.767 0.424 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 66/399 2.07G 0.04193 0.03947 0.004551 71 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.40it/s] all 128 993 0.82 0.679 0.768 0.419 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 67/399 2.07G 0.04159 0.03896 0.003898 134 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.72it/s] all 128 993 0.81 0.702 0.763 0.419 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 68/399 2.07G 0.04193 0.03722 0.004147 175 416: 100% 28/28 [00:05<00:00, 5.07it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.83 0.681 0.761 0.431 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 69/399 2.07G 0.04344 0.03586 0.004153 109 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.59it/s] all 128 993 0.829 0.666 0.761 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 70/399 2.07G 0.04323 0.03889 0.004465 153 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.8 0.711 0.775 0.409 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 71/399 2.07G 0.04161 0.03782 0.004126 221 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.49it/s] all 128 993 0.734 0.716 0.757 0.415 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 72/399 2.07G 0.04219 0.03724 0.003704 166 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.58it/s] all 128 993 0.754 0.715 0.763 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 73/399 2.07G 0.04158 0.03665 0.003915 132 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.72it/s] all 128 993 0.786 0.689 0.777 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 74/399 2.07G 0.04092 0.03602 0.003652 186 416: 100% 28/28 [00:05<00:00, 5.02it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.822 0.697 0.766 0.425 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 75/399 2.07G 0.04149 0.03723 0.003837 143 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.54it/s] all 128 993 0.778 0.704 0.771 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 76/399 2.07G 0.04162 0.03756 0.003525 191 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.81 0.728 0.8 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 77/399 2.07G 0.04177 0.03899 0.003651 108 416: 100% 28/28 [00:07<00:00, 3.64it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.798 0.744 0.794 0.446 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 78/399 2.07G 0.03888 0.03613 0.003468 143 416: 100% 28/28 [00:05<00:00, 5.38it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.44it/s] all 128 993 0.789 0.745 0.781 0.444 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 79/399 2.07G 0.04029 0.03557 0.003603 131 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.773 0.731 0.777 0.44 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 80/399 2.07G 0.04043 0.03565 0.003742 110 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.806 0.716 0.782 0.438 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 81/399 2.07G 0.04181 0.0375 0.003583 204 416: 100% 28/28 [00:05<00:00, 5.05it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.826 0.677 0.767 0.425 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 82/399 2.07G 0.04066 0.03606 0.003646 199 416: 100% 28/28 [00:05<00:00, 5.30it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.813 0.689 0.764 0.432 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 83/399 2.07G 0.04104 0.03741 0.003257 203 416: 100% 28/28 [00:07<00:00, 3.73it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.80it/s] all 128 993 0.828 0.694 0.772 0.443 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 84/399 2.07G 0.04066 0.0366 0.004209 172 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.821 0.713 0.77 0.449 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 85/399 2.07G 0.03907 0.03463 0.00398 225 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.817 0.706 0.785 0.454 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 86/399 2.07G 0.03953 0.0348 0.003737 118 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.56it/s] all 128 993 0.775 0.715 0.758 0.43 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 87/399 2.07G 0.04078 0.0375 0.003581 133 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.762 0.7 0.743 0.42 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 88/399 2.07G 0.03992 0.0362 0.003162 94 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.789 0.736 0.778 0.442 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 89/399 2.07G 0.04049 0.03754 0.003306 126 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.55it/s] all 128 993 0.82 0.718 0.78 0.441 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 90/399 2.07G 0.03979 0.03483 0.003451 191 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.787 0.723 0.763 0.434 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 91/399 2.07G 0.0387 0.03578 0.003148 176 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.81it/s] all 128 993 0.792 0.712 0.785 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 92/399 2.07G 0.03972 0.03634 0.003046 98 416: 100% 28/28 [00:05<00:00, 5.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.812 0.719 0.779 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 93/399 2.07G 0.04086 0.03638 0.003289 108 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.885 0.683 0.789 0.448 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 94/399 2.07G 0.03991 0.03589 0.002949 164 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.78 0.731 0.796 0.448 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 95/399 2.07G 0.0393 0.03544 0.002908 145 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.80it/s] all 128 993 0.851 0.686 0.77 0.439 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 96/399 2.07G 0.03824 0.03485 0.002963 171 416: 100% 28/28 [00:05<00:00, 5.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.794 0.673 0.756 0.417 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 97/399 2.07G 0.03884 0.03586 0.003561 163 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.826 0.67 0.757 0.423 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 98/399 2.07G 0.03871 0.03501 0.003374 135 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.64it/s] all 128 993 0.817 0.71 0.77 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 99/399 2.07G 0.03944 0.03579 0.00279 122 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.788 0.71 0.761 0.434 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 100/399 2.07G 0.03957 0.03657 0.002968 140 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.787 0.734 0.775 0.437 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 101/399 2.07G 0.03874 0.0348 0.003298 115 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.865 0.701 0.784 0.446 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 102/399 2.07G 0.03857 0.03597 0.003032 139 416: 100% 28/28 [00:05<00:00, 5.28it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.794 0.752 0.796 0.452 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 103/399 2.07G 0.03881 0.03726 0.003279 141 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.81it/s] all 128 993 0.813 0.708 0.791 0.444 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 104/399 2.07G 0.03918 0.03588 0.002338 122 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.50it/s] all 128 993 0.801 0.706 0.777 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 105/399 2.07G 0.0391 0.03629 0.002873 73 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.84 0.693 0.776 0.429 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 106/399 2.07G 0.0398 0.03512 0.002705 154 416: 100% 28/28 [00:05<00:00, 5.00it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.33it/s] all 128 993 0.866 0.679 0.778 0.437 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 107/399 2.07G 0.03847 0.03462 0.002753 147 416: 100% 28/28 [00:06<00:00, 4.53it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.804 0.714 0.787 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 108/399 2.07G 0.03896 0.03564 0.00341 142 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.839 0.651 0.774 0.434 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 109/399 2.07G 0.03713 0.03558 0.003284 121 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.742 0.689 0.741 0.41 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 110/399 2.07G 0.0387 0.03593 0.002603 163 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.788 0.685 0.744 0.434 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 111/399 2.07G 0.0379 0.03575 0.002887 147 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.804 0.728 0.784 0.441 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 112/399 2.07G 0.03732 0.03448 0.00275 106 416: 100% 28/28 [00:07<00:00, 3.95it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.845 0.679 0.759 0.439 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 113/399 2.07G 0.03698 0.03488 0.002519 116 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.54it/s] all 128 993 0.808 0.689 0.753 0.439 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 114/399 2.07G 0.03767 0.03505 0.002469 158 416: 100% 28/28 [00:05<00:00, 5.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.767 0.707 0.764 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 115/399 2.07G 0.03868 0.03577 0.002822 239 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.20it/s] all 128 993 0.81 0.687 0.757 0.432 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 116/399 2.07G 0.03681 0.03576 0.002559 91 416: 100% 28/28 [00:05<00:00, 5.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.82it/s] all 128 993 0.793 0.707 0.764 0.446 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 117/399 2.07G 0.03774 0.03383 0.002919 120 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.47it/s] all 128 993 0.839 0.684 0.767 0.439 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 118/399 2.07G 0.03852 0.03242 0.002821 120 416: 100% 28/28 [00:05<00:00, 5.38it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.842 0.675 0.765 0.433 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 119/399 2.07G 0.03686 0.03462 0.002578 118 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.56it/s] all 128 993 0.81 0.703 0.767 0.454 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 120/399 2.07G 0.0385 0.03608 0.002739 150 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.8 0.709 0.773 0.438 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 121/399 2.07G 0.03689 0.03476 0.002589 165 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.807 0.714 0.774 0.444 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 122/399 2.07G 0.03725 0.03331 0.002444 104 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.809 0.714 0.767 0.437 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 123/399 2.07G 0.0371 0.03447 0.00265 139 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.834 0.707 0.778 0.451 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 124/399 2.07G 0.03653 0.03379 0.002381 118 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.56it/s] all 128 993 0.802 0.71 0.767 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 125/399 2.07G 0.03603 0.03361 0.002261 89 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.82 0.721 0.769 0.446 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 126/399 2.07G 0.03673 0.034 0.002161 191 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.828 0.705 0.772 0.438 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 127/399 2.07G 0.03678 0.03145 0.002283 147 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.823 0.721 0.773 0.44 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 128/399 2.07G 0.03665 0.03326 0.002448 143 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.65it/s] all 128 993 0.837 0.698 0.756 0.431 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 129/399 2.07G 0.03735 0.0348 0.002752 123 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.48it/s] all 128 993 0.867 0.707 0.794 0.457 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 130/399 2.07G 0.0372 0.03481 0.002379 190 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.868 0.703 0.794 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 131/399 2.07G 0.03687 0.03439 0.002172 167 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.789 0.728 0.777 0.46 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 132/399 2.07G 0.03772 0.03488 0.002439 134 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.804 0.713 0.778 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 133/399 2.07G 0.03678 0.03475 0.002442 123 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.795 0.689 0.768 0.445 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 134/399 2.07G 0.03608 0.03312 0.002737 157 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.792 0.722 0.772 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 135/399 2.07G 0.03695 0.03587 0.002361 112 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.858 0.674 0.778 0.457 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 136/399 2.07G 0.03657 0.03391 0.002295 175 416: 100% 28/28 [00:07<00:00, 3.91it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.56it/s] all 128 993 0.82 0.712 0.782 0.456 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 137/399 2.07G 0.03539 0.03452 0.002045 138 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.70it/s] all 128 993 0.847 0.696 0.767 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 138/399 2.07G 0.03648 0.03528 0.002083 148 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.57it/s] all 128 993 0.825 0.71 0.779 0.446 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 139/399 2.07G 0.0354 0.03354 0.002459 176 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.78it/s] all 128 993 0.822 0.707 0.788 0.454 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 140/399 2.07G 0.0365 0.03324 0.002237 157 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.809 0.72 0.791 0.448 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 141/399 2.07G 0.036 0.03566 0.001972 240 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.12it/s] all 128 993 0.79 0.741 0.796 0.452 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 142/399 2.07G 0.03682 0.03272 0.002136 138 416: 100% 28/28 [00:06<00:00, 4.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.807 0.721 0.78 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 143/399 2.07G 0.03546 0.03229 0.002085 110 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.58it/s] all 128 993 0.823 0.719 0.776 0.456 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 144/399 2.07G 0.03584 0.03201 0.002387 142 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.805 0.722 0.785 0.458 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 145/399 2.07G 0.03606 0.03394 0.002262 109 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.65it/s] all 128 993 0.778 0.728 0.772 0.454 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 146/399 2.07G 0.03583 0.03456 0.00199 165 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.752 0.74 0.769 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 147/399 2.07G 0.03551 0.03552 0.002483 118 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.801 0.708 0.758 0.443 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 148/399 2.07G 0.03488 0.03179 0.002141 145 416: 100% 28/28 [00:05<00:00, 5.30it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.832 0.668 0.751 0.452 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 149/399 2.07G 0.03475 0.03281 0.002621 94 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.81 0.729 0.783 0.457 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 150/399 2.07G 0.03606 0.03473 0.002194 209 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.39it/s] all 128 993 0.819 0.72 0.778 0.447 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 151/399 2.07G 0.0356 0.03335 0.002023 120 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.851 0.721 0.781 0.458 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 152/399 2.07G 0.03554 0.03345 0.002285 262 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.849 0.698 0.773 0.449 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 153/399 2.07G 0.0366 0.03422 0.002163 169 416: 100% 28/28 [00:05<00:00, 5.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.805 0.71 0.771 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 154/399 2.07G 0.03455 0.03187 0.002106 78 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.809 0.732 0.772 0.457 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 155/399 2.07G 0.03503 0.03271 0.001853 167 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.72it/s] all 128 993 0.815 0.731 0.778 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 156/399 2.07G 0.03466 0.03345 0.002204 114 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.863 0.693 0.781 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 157/399 2.07G 0.03522 0.03391 0.002453 113 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.37it/s] all 128 993 0.848 0.689 0.778 0.458 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 158/399 2.07G 0.0347 0.03259 0.002256 105 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.81it/s] all 128 993 0.818 0.694 0.768 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 159/399 2.07G 0.03536 0.03404 0.002177 192 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.70it/s] all 128 993 0.814 0.702 0.766 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 160/399 2.07G 0.03464 0.0321 0.002089 216 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.35it/s] all 128 993 0.845 0.703 0.775 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 161/399 2.07G 0.03497 0.0327 0.002164 152 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.805 0.752 0.801 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 162/399 2.07G 0.0351 0.03567 0.001816 143 416: 100% 28/28 [00:05<00:00, 4.97it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.50it/s] all 128 993 0.821 0.708 0.779 0.455 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 163/399 2.07G 0.03568 0.03216 0.001854 175 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.41it/s] all 128 993 0.821 0.698 0.78 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 164/399 2.07G 0.0338 0.03055 0.002179 85 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.59it/s] all 128 993 0.819 0.711 0.776 0.456 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 165/399 2.07G 0.03415 0.03332 0.002087 126 416: 100% 28/28 [00:06<00:00, 4.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.44it/s] all 128 993 0.813 0.707 0.778 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 166/399 2.07G 0.03487 0.03277 0.002013 134 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.816 0.725 0.793 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 167/399 2.07G 0.03322 0.03053 0.001954 155 416: 100% 28/28 [00:05<00:00, 5.30it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.42it/s] all 128 993 0.86 0.702 0.793 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 168/399 2.07G 0.03519 0.03171 0.001957 136 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.822 0.74 0.786 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 169/399 2.07G 0.03425 0.03137 0.001857 178 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.58it/s] all 128 993 0.844 0.702 0.788 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 170/399 2.07G 0.03386 0.0316 0.002072 145 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.795 0.749 0.791 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 171/399 2.07G 0.03347 0.03128 0.001923 115 416: 100% 28/28 [00:07<00:00, 3.87it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.87it/s] all 128 993 0.842 0.725 0.79 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 172/399 2.07G 0.03423 0.03321 0.001802 196 416: 100% 28/28 [00:05<00:00, 5.04it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.90it/s] all 128 993 0.84 0.73 0.792 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 173/399 2.07G 0.03456 0.03313 0.001641 108 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.813 0.743 0.787 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 174/399 2.07G 0.03319 0.03204 0.001883 147 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.78it/s] all 128 993 0.809 0.727 0.77 0.452 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 175/399 2.07G 0.03356 0.03251 0.001745 193 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.835 0.712 0.776 0.45 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 176/399 2.07G 0.03421 0.03421 0.001972 157 416: 100% 28/28 [00:05<00:00, 5.03it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.863 0.703 0.779 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 177/399 2.07G 0.03293 0.03045 0.001742 157 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.834 0.698 0.774 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 178/399 2.07G 0.03402 0.03165 0.001847 160 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.842 0.708 0.766 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 179/399 2.07G 0.03414 0.031 0.001673 128 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.825 0.724 0.766 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 180/399 2.07G 0.0336 0.0325 0.001998 197 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.855 0.714 0.769 0.458 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 181/399 2.07G 0.03262 0.03001 0.00155 140 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.828 0.737 0.781 0.475 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 182/399 2.07G 0.03389 0.03417 0.001697 119 416: 100% 28/28 [00:05<00:00, 4.97it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.64it/s] all 128 993 0.832 0.727 0.775 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 183/399 2.07G 0.03376 0.03097 0.001785 110 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.82 0.754 0.783 0.458 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 184/399 2.07G 0.0337 0.03181 0.001714 231 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.804 0.721 0.775 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 185/399 2.07G 0.03285 0.03143 0.001814 151 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.801 0.747 0.786 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 186/399 2.07G 0.03343 0.0321 0.001603 85 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.37it/s] all 128 993 0.837 0.718 0.789 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 187/399 2.07G 0.03248 0.03019 0.001573 124 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.829 0.723 0.785 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 188/399 2.07G 0.03379 0.03308 0.001565 120 416: 100% 28/28 [00:05<00:00, 5.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.822 0.717 0.79 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 189/399 2.07G 0.03334 0.03284 0.001437 185 416: 100% 28/28 [00:05<00:00, 4.99it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.812 0.752 0.794 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 190/399 2.07G 0.03335 0.03131 0.001649 136 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.70it/s] all 128 993 0.824 0.708 0.78 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 191/399 2.07G 0.03333 0.03159 0.001567 104 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.812 0.72 0.785 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 192/399 2.07G 0.03214 0.03046 0.001885 134 416: 100% 28/28 [00:05<00:00, 5.32it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.57it/s] all 128 993 0.844 0.715 0.793 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 193/399 2.07G 0.03332 0.03161 0.001591 124 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.84 0.737 0.798 0.486 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 194/399 2.07G 0.03281 0.03294 0.001636 140 416: 100% 28/28 [00:05<00:00, 5.04it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.55it/s] all 128 993 0.795 0.767 0.795 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 195/399 2.07G 0.03377 0.03089 0.001516 92 416: 100% 28/28 [00:06<00:00, 4.34it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.86it/s] all 128 993 0.858 0.702 0.779 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 196/399 2.07G 0.03324 0.03145 0.001729 203 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.846 0.709 0.788 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 197/399 2.07G 0.0328 0.03184 0.001925 105 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.22it/s] all 128 993 0.836 0.726 0.778 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 198/399 2.07G 0.03282 0.02993 0.002147 152 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.864 0.699 0.775 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 199/399 2.07G 0.03149 0.03104 0.00177 134 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.74it/s] all 128 993 0.879 0.693 0.777 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 200/399 2.07G 0.03251 0.03237 0.001452 146 416: 100% 28/28 [00:06<00:00, 4.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.45it/s] all 128 993 0.847 0.717 0.786 0.474 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 201/399 2.07G 0.03256 0.02993 0.0013 239 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.81it/s] all 128 993 0.867 0.731 0.775 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 202/399 2.07G 0.03256 0.0318 0.001514 227 416: 100% 28/28 [00:05<00:00, 5.02it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.94it/s] all 128 993 0.855 0.722 0.781 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 203/399 2.07G 0.03222 0.03066 0.001636 103 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.76it/s] all 128 993 0.856 0.714 0.778 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 204/399 2.07G 0.03131 0.03004 0.001346 139 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.856 0.707 0.775 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 205/399 2.07G 0.03317 0.03064 0.001753 180 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.56it/s] all 128 993 0.842 0.69 0.771 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 206/399 2.07G 0.03269 0.03014 0.001744 222 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.58it/s] all 128 993 0.819 0.732 0.782 0.46 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 207/399 2.07G 0.03171 0.03166 0.001503 190 416: 100% 28/28 [00:05<00:00, 5.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.80it/s] all 128 993 0.835 0.714 0.791 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 208/399 2.07G 0.03195 0.03165 0.001394 157 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.831 0.721 0.786 0.473 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 209/399 2.07G 0.03263 0.03084 0.001543 195 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.59it/s] all 128 993 0.821 0.717 0.778 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 210/399 2.07G 0.03279 0.03152 0.001594 181 416: 100% 28/28 [00:05<00:00, 5.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.78it/s] all 128 993 0.827 0.707 0.778 0.46 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 211/399 2.07G 0.03266 0.03153 0.001746 132 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.795 0.731 0.78 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 212/399 2.07G 0.03132 0.03009 0.001488 140 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.51it/s] all 128 993 0.811 0.722 0.777 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 213/399 2.07G 0.03143 0.02845 0.001509 118 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.823 0.726 0.772 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 214/399 2.07G 0.03166 0.02946 0.001359 101 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.848 0.7 0.774 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 215/399 2.07G 0.03118 0.03051 0.001387 160 416: 100% 28/28 [00:05<00:00, 5.09it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.826 0.716 0.789 0.476 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 216/399 2.07G 0.03139 0.03092 0.001583 176 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.791 0.733 0.791 0.475 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 217/399 2.07G 0.03081 0.02945 0.001512 166 416: 100% 28/28 [00:05<00:00, 5.23it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.819 0.706 0.781 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 218/399 2.07G 0.0309 0.03083 0.001387 205 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.838 0.71 0.786 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 219/399 2.07G 0.03039 0.02911 0.001589 138 416: 100% 28/28 [00:05<00:00, 5.31it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.67it/s] all 128 993 0.817 0.718 0.79 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 220/399 2.07G 0.03194 0.0294 0.001416 122 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.61it/s] all 128 993 0.833 0.697 0.784 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 221/399 2.07G 0.03127 0.03031 0.001331 116 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.809 0.722 0.788 0.472 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 222/399 2.07G 0.03151 0.03041 0.00152 117 416: 100% 28/28 [00:05<00:00, 5.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.833 0.718 0.783 0.46 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 223/399 2.07G 0.03154 0.02827 0.001512 117 416: 100% 28/28 [00:05<00:00, 5.26it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.835 0.734 0.786 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 224/399 2.07G 0.03088 0.02865 0.00131 119 416: 100% 28/28 [00:06<00:00, 4.61it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.24it/s] all 128 993 0.83 0.731 0.788 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 225/399 2.07G 0.0309 0.02972 0.001396 109 416: 100% 28/28 [00:05<00:00, 4.81it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.848 0.728 0.78 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 226/399 2.07G 0.03158 0.03111 0.001468 93 416: 100% 28/28 [00:05<00:00, 4.98it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.87it/s] all 128 993 0.841 0.748 0.79 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 227/399 2.07G 0.03074 0.02914 0.001209 99 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.88it/s] all 128 993 0.858 0.723 0.79 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 228/399 2.07G 0.03091 0.02952 0.00133 175 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.65it/s] all 128 993 0.865 0.718 0.796 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 229/399 2.07G 0.03078 0.02893 0.001462 161 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.85it/s] all 128 993 0.857 0.729 0.792 0.476 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 230/399 2.07G 0.03068 0.02809 0.001456 130 416: 100% 28/28 [00:06<00:00, 4.21it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.83it/s] all 128 993 0.848 0.739 0.801 0.472 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 231/399 2.07G 0.03076 0.03136 0.00129 177 416: 100% 28/28 [00:05<00:00, 4.96it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.87it/s] all 128 993 0.81 0.741 0.795 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 232/399 2.07G 0.03099 0.02965 0.001283 146 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.35it/s] all 128 993 0.835 0.726 0.791 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 233/399 2.07G 0.03015 0.02999 0.001274 123 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.835 0.731 0.783 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 234/399 2.07G 0.03032 0.03088 0.001295 153 416: 100% 28/28 [00:05<00:00, 5.17it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.19it/s] all 128 993 0.812 0.732 0.777 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 235/399 2.07G 0.03134 0.03063 0.001338 106 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.46it/s] all 128 993 0.846 0.718 0.786 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 236/399 2.07G 0.03155 0.02958 0.001764 139 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.70it/s] all 128 993 0.824 0.748 0.78 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 237/399 2.07G 0.0305 0.02995 0.001175 150 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.68it/s] all 128 993 0.837 0.703 0.772 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 238/399 2.07G 0.03156 0.03018 0.00114 151 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.851 0.709 0.781 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 239/399 2.07G 0.02986 0.02924 0.001164 105 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.82it/s] all 128 993 0.838 0.716 0.772 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 240/399 2.07G 0.03014 0.0293 0.001432 198 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.814 0.728 0.775 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 241/399 2.07G 0.03064 0.02858 0.001351 141 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.71it/s] all 128 993 0.829 0.717 0.769 0.46 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 242/399 2.07G 0.02954 0.02689 0.001357 90 416: 100% 28/28 [00:05<00:00, 5.32it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.856 0.694 0.774 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 243/399 2.07G 0.02977 0.02808 0.001349 125 416: 100% 28/28 [00:05<00:00, 5.01it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.79it/s] all 128 993 0.805 0.711 0.767 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 244/399 2.07G 0.03058 0.03029 0.001373 170 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.866 0.709 0.776 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 245/399 2.07G 0.03018 0.02731 0.001257 157 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.864 0.694 0.775 0.459 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 246/399 2.07G 0.03023 0.02914 0.001216 171 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.863 0.71 0.777 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 247/399 2.07G 0.03005 0.02952 0.001298 107 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.849 0.69 0.776 0.453 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 248/399 2.07G 0.03 0.02823 0.001636 141 416: 100% 28/28 [00:05<00:00, 5.18it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.842 0.699 0.779 0.456 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 249/399 2.07G 0.02965 0.02943 0.001261 151 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.81it/s] all 128 993 0.84 0.725 0.78 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 250/399 2.07G 0.02988 0.02882 0.001237 173 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.60it/s] all 128 993 0.881 0.71 0.78 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 251/399 2.07G 0.02966 0.02872 0.001369 169 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.835 0.739 0.782 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 252/399 2.07G 0.03007 0.02848 0.001293 151 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.82it/s] all 128 993 0.847 0.735 0.783 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 253/399 2.07G 0.03078 0.03021 0.00124 100 416: 100% 28/28 [00:05<00:00, 5.04it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.843 0.719 0.774 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 254/399 2.07G 0.03051 0.02809 0.001102 115 416: 100% 28/28 [00:05<00:00, 5.29it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.855 0.718 0.779 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 255/399 2.07G 0.03036 0.03011 0.001298 146 416: 100% 28/28 [00:07<00:00, 3.87it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.07it/s] all 128 993 0.844 0.724 0.784 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 256/399 2.07G 0.03065 0.03066 0.001471 126 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.91it/s] all 128 993 0.801 0.728 0.779 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 257/399 2.07G 0.02965 0.02793 0.001255 103 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.82 0.731 0.774 0.462 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 258/399 2.07G 0.02863 0.0277 0.00103 143 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.83it/s] all 128 993 0.821 0.724 0.772 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 259/399 2.07G 0.02942 0.02817 0.001086 166 416: 100% 28/28 [00:06<00:00, 4.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.34it/s] all 128 993 0.821 0.737 0.776 0.465 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 260/399 2.07G 0.03019 0.02991 0.001212 110 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.74it/s] all 128 993 0.837 0.734 0.789 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 261/399 2.07G 0.02989 0.02901 0.001357 195 416: 100% 28/28 [00:05<00:00, 5.08it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.831 0.746 0.783 0.473 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 262/399 2.07G 0.02943 0.02971 0.001151 158 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.52it/s] all 128 993 0.841 0.723 0.783 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 263/399 2.07G 0.02863 0.02751 0.001289 143 416: 100% 28/28 [00:05<00:00, 5.40it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.72it/s] all 128 993 0.856 0.716 0.78 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 264/399 2.07G 0.02938 0.02985 0.001068 192 416: 100% 28/28 [00:05<00:00, 5.07it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.89it/s] all 128 993 0.857 0.718 0.79 0.476 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 265/399 2.07G 0.02907 0.02819 0.001378 111 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.843 0.729 0.791 0.473 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 266/399 2.07G 0.02907 0.02908 0.001412 204 416: 100% 28/28 [00:05<00:00, 5.12it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.62it/s] all 128 993 0.874 0.693 0.787 0.473 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 267/399 2.07G 0.03057 0.02958 0.00103 170 416: 100% 28/28 [00:05<00:00, 5.07it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.72it/s] all 128 993 0.825 0.716 0.774 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 268/399 2.07G 0.02942 0.02792 0.001217 168 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.86it/s] all 128 993 0.825 0.723 0.777 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 269/399 2.07G 0.02951 0.02877 0.001179 178 416: 100% 28/28 [00:05<00:00, 5.24it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.15it/s] all 128 993 0.848 0.711 0.793 0.478 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 270/399 2.07G 0.02873 0.02931 0.0008756 225 416: 100% 28/28 [00:05<00:00, 5.05it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.796 0.756 0.794 0.48 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 271/399 2.07G 0.02894 0.02899 0.001071 173 416: 100% 28/28 [00:05<00:00, 5.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.83it/s] all 128 993 0.84 0.729 0.795 0.474 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 272/399 2.07G 0.02901 0.02816 0.001224 168 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.85it/s] all 128 993 0.822 0.734 0.786 0.474 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 273/399 2.07G 0.02905 0.02812 0.001156 107 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.822 0.743 0.784 0.473 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 274/399 2.07G 0.02978 0.02815 0.001496 106 416: 100% 28/28 [00:05<00:00, 5.30it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.57it/s] all 128 993 0.865 0.712 0.784 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 275/399 2.07G 0.02864 0.02703 0.001142 176 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.57it/s] all 128 993 0.86 0.705 0.792 0.463 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 276/399 2.07G 0.02889 0.02795 0.0009889 189 416: 100% 28/28 [00:05<00:00, 5.16it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.29it/s] all 128 993 0.824 0.723 0.795 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 277/399 2.07G 0.02926 0.02794 0.001263 137 416: 100% 28/28 [00:05<00:00, 5.10it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.82it/s] all 128 993 0.832 0.73 0.788 0.461 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 278/399 2.07G 0.02906 0.02995 0.001215 155 416: 100% 28/28 [00:05<00:00, 5.06it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.53it/s] all 128 993 0.848 0.723 0.79 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 279/399 2.07G 0.02875 0.02937 0.001181 219 416: 100% 28/28 [00:05<00:00, 5.02it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.856 0.715 0.782 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 280/399 2.07G 0.02888 0.02932 0.0008666 176 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.47it/s] all 128 993 0.851 0.719 0.782 0.467 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 281/399 2.07G 0.02831 0.02866 0.000874 147 416: 100% 28/28 [00:05<00:00, 5.13it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.869 0.725 0.782 0.464 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 282/399 2.07G 0.02912 0.02812 0.001166 142 416: 100% 28/28 [00:05<00:00, 5.25it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.69it/s] all 128 993 0.837 0.744 0.785 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 283/399 2.07G 0.02788 0.02685 0.001342 144 416: 100% 28/28 [00:05<00:00, 5.15it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.93it/s] all 128 993 0.852 0.722 0.791 0.468 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 284/399 2.07G 0.02819 0.02677 0.001044 137 416: 100% 28/28 [00:05<00:00, 5.19it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.75it/s] all 128 993 0.857 0.707 0.791 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 285/399 2.07G 0.02913 0.02849 0.001269 116 416: 100% 28/28 [00:07<00:00, 3.82it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.66it/s] all 128 993 0.831 0.724 0.786 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 286/399 2.07G 0.0289 0.02717 0.0009714 168 416: 100% 28/28 [00:05<00:00, 5.27it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.59it/s] all 128 993 0.842 0.722 0.785 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 287/399 2.07G 0.02837 0.02693 0.0009378 151 416: 100% 28/28 [00:05<00:00, 5.20it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.86it/s] all 128 993 0.849 0.731 0.786 0.466 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 288/399 2.07G 0.02849 0.02804 0.0009775 130 416: 100% 28/28 [00:05<00:00, 5.11it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:01<00:00, 3.38it/s] all 128 993 0.869 0.726 0.787 0.47 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 289/399 2.07G 0.02936 0.02761 0.001281 171 416: 100% 28/28 [00:06<00:00, 4.46it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.84it/s] all 128 993 0.837 0.73 0.786 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 290/399 2.07G 0.02877 0.02791 0.0009865 133 416: 100% 28/28 [00:05<00:00, 5.22it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.63it/s] all 128 993 0.86 0.725 0.782 0.469 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 291/399 2.07G 0.02853 0.02726 0.00117 128 416: 100% 28/28 [00:05<00:00, 5.14it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.876 0.715 0.79 0.471 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 292/399 2.07G 0.02882 0.02785 0.001067 192 416: 100% 28/28 [00:05<00:00, 5.04it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.77it/s] all 128 993 0.857 0.714 0.789 0.475 Epoch GPU_mem box_loss obj_loss cls_loss Instances Size 293/399 2.07G 0.02902 0.02935 0.001364 172 416: 100% 28/28 [00:05<00:00, 5.01it/s] Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:00<00:00, 4.73it/s] all 128 993 0.859 0.712 0.794 0.479 Stopping training early as no improvement observed in last 100 epochs. Best results observed at epoch 193, best model saved as best.pt. To update EarlyStopping(patience=100) pass a new patience value, i.e. `python train.py --patience 300` or use `--patience 0` to disable EarlyStopping. 294 epochs completed in 0.559 hours. Optimizer stripped from runs/train/exp/weights/last.pt, 14.3MB Optimizer stripped from runs/train/exp/weights/best.pt, 14.3MB Validating runs/train/exp/weights/best.pt... Fusing layers... Model summary: 157 layers, 7029004 parameters, 0 gradients, 15.8 GFLOPs Class Images Instances P R mAP50 mAP50-95: 100% 4/4 [00:03<00:00, 1.14it/s] all 128 993 0.842 0.738 0.798 0.486 fish 128 504 0.807 0.656 0.723 0.385 jellyfish 128 226 0.878 0.796 0.869 0.534 penguin 128 86 0.872 0.767 0.846 0.434 puffin 128 51 0.777 0.667 0.64 0.34 shark 128 67 0.784 0.701 0.805 0.494 starfish 128 20 0.898 0.883 0.934 0.708 stingray 128 39 0.876 0.692 0.77 0.505 Results saved to runs/train/exp
Evaluate Custom YOLOv5 Detector Performance
Using tensorboard we are going to evaluate our model.
In [ ]:
# Start tensorboard
# Launch after you have started training
# logs save in the folder "runs"
%load_ext tensorboard
%tensorboard --logdir runs
In [ ]:
# Run inference with trained weights
!python detect.py --weights runs/train/exp/weights/best.pt --img 416 --conf 0.1 --source {dataset.location}/test/images
detect: weights=['runs/train/exp/weights/best.pt'], source=/content/datasets/fish-1/test/images, data=data/coco128.yaml, imgsz=[416, 416], conf_thres=0.1, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=runs/detect, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False, vid_stride=1 YOLOv5 🚀 v7.0-71-gc442a2e Python-3.8.10 torch-1.13.1+cu116 CUDA:0 (Tesla T4, 15110MiB) Fusing layers... Model summary: 157 layers, 7029004 parameters, 0 gradients, 15.8 GFLOPs image 1/63 /content/datasets/fish-1/test/images/IMG_2298_jpeg_jpg.rf.6fc31257f43a6f3bb3c4f7e44bb966d0.jpg: 416x416 8 puffins, 8.3ms image 2/63 /content/datasets/fish-1/test/images/IMG_2304_jpeg_jpg.rf.036da64118c64326b72acfc230a0b48e.jpg: 416x416 14 penguins, 9.2ms image 3/63 /content/datasets/fish-1/test/images/IMG_2311_jpeg_jpg.rf.68f11e5acd510450caa3b09142ebd318.jpg: 416x416 1 penguin, 8.8ms image 4/63 /content/datasets/fish-1/test/images/IMG_2322_jpeg_jpg.rf.d953b6f4edb65fc8f46b32544b9b108c.jpg: 416x416 23 penguins, 8.9ms image 5/63 /content/datasets/fish-1/test/images/IMG_2352_jpeg_jpg.rf.e35cf0992ba07451296021e9852558c9.jpg: 416x416 3 penguins, 1 starfish, 8.8ms image 6/63 /content/datasets/fish-1/test/images/IMG_2370_jpeg_jpg.rf.1b601a38384256840af13d4bf5ae2278.jpg: 416x416 4 fishs, 8.4ms image 7/63 /content/datasets/fish-1/test/images/IMG_2373_jpeg_jpg.rf.e8aba66244ca589f64745cf0ed674edb.jpg: 416x416 1 fish, 1 starfish, 8.9ms image 8/63 /content/datasets/fish-1/test/images/IMG_2375_jpeg_jpg.rf.28bcb7eba2f3c23c3d87f52167223325.jpg: 416x416 1 fish, 8.9ms image 9/63 /content/datasets/fish-1/test/images/IMG_2384_jpeg_jpg.rf.75dd4d152d6aac33b47f7bcea6d884dd.jpg: 416x416 1 fish, 1 starfish, 8.4ms image 10/63 /content/datasets/fish-1/test/images/IMG_2405_jpeg_jpg.rf.824ad31aabfe8760577ae739e0e76904.jpg: 416x416 6 fishs, 8.2ms image 11/63 /content/datasets/fish-1/test/images/IMG_2415_jpeg_jpg.rf.71994e70c563aafc4bbcacb9ecbe8435.jpg: 416x416 19 fishs, 1 shark, 8.2ms image 12/63 /content/datasets/fish-1/test/images/IMG_2418_jpeg_jpg.rf.b9f7491c70dd5577a609f0122670038a.jpg: 416x416 23 fishs, 2 sharks, 2 stingrays, 8.2ms image 13/63 /content/datasets/fish-1/test/images/IMG_2435_jpeg_jpg.rf.40de207629bf0552c1f398f3ae4ce06c.jpg: 416x416 8 fishs, 1 shark, 8.2ms image 14/63 /content/datasets/fish-1/test/images/IMG_2444_jpeg_jpg.rf.b56b24896534111a3245e963c4e8cd3e.jpg: 416x416 14 fishs, 4 sharks, 8.7ms image 15/63 /content/datasets/fish-1/test/images/IMG_2449_jpeg_jpg.rf.17e393c419d57ad5d57d72f4728735c7.jpg: 416x416 19 fishs, 5 sharks, 8.4ms image 16/63 /content/datasets/fish-1/test/images/IMG_2469_jpeg_jpg.rf.c600f69bf682818937703ef5729a3155.jpg: 416x416 21 jellyfishs, 9.2ms image 17/63 /content/datasets/fish-1/test/images/IMG_2494_jpeg_jpg.rf.4beb1d6ba29c67e0c9f1629ae00267e2.jpg: 416x416 25 fishs, 5 sharks, 1 stingray, 8.6ms image 18/63 /content/datasets/fish-1/test/images/IMG_2504_jpeg_jpg.rf.862b2ea8301eddcbdb3e168bd3a536dd.jpg: 416x416 13 fishs, 3 sharks, 8.2ms image 19/63 /content/datasets/fish-1/test/images/IMG_2517_jpeg_jpg.rf.1dcdfb92d458d632b95bea285c4d29e1.jpg: 416x416 1 fish, 1 shark, 1 stingray, 8.2ms image 20/63 /content/datasets/fish-1/test/images/IMG_2523_jpeg_jpg.rf.2de7d47742dc5c0da171efedc7503110.jpg: 416x416 1 fish, 26 puffins, 8.1ms image 21/63 /content/datasets/fish-1/test/images/IMG_2532_jpeg_jpg.rf.0451abf9a71fc347ce5175005b3a9a1e.jpg: 416x416 3 starfishs, 8.2ms image 22/63 /content/datasets/fish-1/test/images/IMG_2533_jpeg_jpg.rf.c7904822bfe93389f2131fe7905e18c6.jpg: 416x416 1 fish, 3 starfishs, 8.1ms image 23/63 /content/datasets/fish-1/test/images/IMG_2536_jpeg_jpg.rf.080362df656db5c477f63790206a2453.jpg: 416x416 1 fish, 1 starfish, 8.2ms image 24/63 /content/datasets/fish-1/test/images/IMG_2557_jpeg_jpg.rf.c0cc4e818ce5736c8eeed3e046753a5e.jpg: 416x416 1 stingray, 8.2ms image 25/63 /content/datasets/fish-1/test/images/IMG_2558_jpeg_jpg.rf.65914b818b6895c49863a305d3bf5ec7.jpg: 416x416 12 fishs, 8 sharks, 1 stingray, 8.2ms image 26/63 /content/datasets/fish-1/test/images/IMG_2560_jpeg_jpg.rf.5858c0ecc76a95b079456aa584dc2b33.jpg: 416x416 10 fishs, 3 sharks, 1 stingray, 9.4ms image 27/63 /content/datasets/fish-1/test/images/IMG_2565_jpeg_jpg.rf.5aec66a6cf456177497fc920a8833192.jpg: 416x416 5 fishs, 2 sharks, 1 stingray, 8.2ms image 28/63 /content/datasets/fish-1/test/images/IMG_2579_jpeg_jpg.rf.8f614d492075f6edb32f557cc8273fe1.jpg: 416x416 20 fishs, 4 sharks, 2 stingrays, 8.7ms image 29/63 /content/datasets/fish-1/test/images/IMG_2585_jpeg_jpg.rf.5f32306408fdd760a6233a02f5b5d6bb.jpg: 416x416 3 stingrays, 11.9ms image 30/63 /content/datasets/fish-1/test/images/IMG_2588_jpeg_jpg.rf.56251f92dc3c1e1bad20729eef6cb4af.jpg: 416x416 4 stingrays, 8.3ms image 31/63 /content/datasets/fish-1/test/images/IMG_2593_jpeg_jpg.rf.9dcaddf5d4ae064cbf194b6ab6aefb58.jpg: 416x416 4 fishs, 1 shark, 1 stingray, 8.1ms image 32/63 /content/datasets/fish-1/test/images/IMG_2607_jpeg_jpg.rf.7a6c4a12a93362234b24a2e49e30ea0a.jpg: 416x416 10 fishs, 4 sharks, 1 stingray, 8.2ms image 33/63 /content/datasets/fish-1/test/images/IMG_2620_jpeg_jpg.rf.7184e8514c9b5ed372ffbcf7325c682d.jpg: 416x416 2 fishs, 1 stingray, 8.2ms image 34/63 /content/datasets/fish-1/test/images/IMG_2640_jpeg_jpg.rf.702f9a193d599607b51fbbde2ba3c1ba.jpg: 416x416 1 stingray, 8.1ms image 35/63 /content/datasets/fish-1/test/images/IMG_2655_jpeg_jpg.rf.c06ae257d719766bf0eb261fb280dac5.jpg: 416x416 2 stingrays, 8.2ms image 36/63 /content/datasets/fish-1/test/images/IMG_2657_jpeg_jpg.rf.29c074e1588a80654de22a1a9a1573c3.jpg: 416x416 5 fishs, 1 shark, 9.2ms image 37/63 /content/datasets/fish-1/test/images/IMG_3121_jpeg_jpg.rf.31152be397d63a40c8f2646c2ba78c85.jpg: 416x416 5 starfishs, 9.2ms image 38/63 /content/datasets/fish-1/test/images/IMG_3126_jpeg_jpg.rf.089ab7e7ea3a78eef23cc866fbd81c6c.jpg: 416x416 1 fish, 4 starfishs, 8.2ms image 39/63 /content/datasets/fish-1/test/images/IMG_3134_jpeg_jpg.rf.8494acbce1c29ea685fb2759b6ccd6e4.jpg: 416x416 4 puffins, 8.2ms image 40/63 /content/datasets/fish-1/test/images/IMG_3140_jpeg_jpg.rf.bdc84fbedf9e2a61cf2adbed96bfae21.jpg: 416x416 12 puffins, 8.2ms image 41/63 /content/datasets/fish-1/test/images/IMG_3148_jpeg_jpg.rf.78f3a5e0eb9eb6d4892b913f4d5ac24a.jpg: 416x416 1 puffin, 8.2ms image 42/63 /content/datasets/fish-1/test/images/IMG_3152_jpeg_jpg.rf.640de7373e4d8f8f3dee531cb4f4794d.jpg: 416x416 3 puffins, 8.3ms image 43/63 /content/datasets/fish-1/test/images/IMG_3173_jpeg_jpg.rf.9dd4df5f5709d79c6d4b2497dcf6b38c.jpg: 416x416 10 penguins, 8.8ms image 44/63 /content/datasets/fish-1/test/images/IMG_3178_jpeg_jpg.rf.c3c3e92efab4d5bece997907b780696b.jpg: 416x416 2 fishs, 8.1ms image 45/63 /content/datasets/fish-1/test/images/IMG_3179_jpeg_jpg.rf.c1ae586f7212418351643c14df61fe20.jpg: 416x416 2 fishs, 1 starfish, 8.2ms image 46/63 /content/datasets/fish-1/test/images/IMG_3181_jpeg_jpg.rf.5128770221a6c40ebc883f3859d54ca4.jpg: 416x416 1 fish, 2 starfishs, 8.2ms image 47/63 /content/datasets/fish-1/test/images/IMG_8420_jpg.rf.c2eb246730c16ed27a7933858c7b28fa.jpg: 416x416 37 fishs, 2 sharks, 8.3ms image 48/63 /content/datasets/fish-1/test/images/IMG_8445_jpg.rf.4ee3c8d9343f149e2a1e2a92fdde2dc1.jpg: 416x416 26 fishs, 1 jellyfish, 1 shark, 2 stingrays, 8.2ms image 49/63 /content/datasets/fish-1/test/images/IMG_8502_jpg.rf.29074c14878aaddaf58849668eb70cc7.jpg: 416x416 4 fishs, 8.3ms image 50/63 /content/datasets/fish-1/test/images/IMG_8517_MOV-0_jpg.rf.1fea754eadee7927df4f87c85928d28d.jpg: 416x416 6 fishs, 12.5ms image 51/63 /content/datasets/fish-1/test/images/IMG_8520_jpg.rf.2e20c6217a1af671e8b62549f7e155e7.jpg: 416x416 17 fishs, 2 puffins, 8.6ms image 52/63 /content/datasets/fish-1/test/images/IMG_8525_jpg.rf.0f1e734a56d4c44c48a7e45cc7d89cc6.jpg: 416x416 6 fishs, 8.2ms image 53/63 /content/datasets/fish-1/test/images/IMG_8534_jpg.rf.ded2ba4bb161a7169abd8c4dbdd7971a.jpg: 416x416 9 puffins, 8.2ms image 54/63 /content/datasets/fish-1/test/images/IMG_8535_MOV-1_jpg.rf.2195bcef31a04461c9eb7c32d5134736.jpg: 416x416 9 puffins, 8.8ms image 55/63 /content/datasets/fish-1/test/images/IMG_8536_jpg.rf.dd199338a55810901aa2a999aa36baa3.jpg: 416x416 10 fishs, 8.1ms image 56/63 /content/datasets/fish-1/test/images/IMG_8538_jpg.rf.f071f29d882c5e1460d30993be07d799.jpg: 416x416 8 fishs, 8.2ms image 57/63 /content/datasets/fish-1/test/images/IMG_8545_jpg.rf.9de3b9302da7ce7e9b7a23cf672ee696.jpg: 416x416 18 fishs, 8.2ms image 58/63 /content/datasets/fish-1/test/images/IMG_8551_MOV-2_jpg.rf.c0c7c293c0b08a9c168f9f64445fee8e.jpg: 416x416 6 fishs, 1 penguin, 8.8ms image 59/63 /content/datasets/fish-1/test/images/IMG_8578_MOV-0_jpg.rf.58d1cc91cc140626570bdfb9590b46c5.jpg: 416x416 1 fish, 8.2ms image 60/63 /content/datasets/fish-1/test/images/IMG_8579_jpg.rf.df13aad58398dce547492ac2e0782223.jpg: 416x416 51 fishs, 1 jellyfish, 1 stingray, 8.5ms image 61/63 /content/datasets/fish-1/test/images/IMG_8590_MOV-5_jpg.rf.9c42b0632da35cedc77aef722deec3cf.jpg: 416x416 1 fish, 1 jellyfish, 1 stingray, 8.2ms image 62/63 /content/datasets/fish-1/test/images/IMG_8591_MOV-1_jpg.rf.7223fe0cbf72f6806b4f6e3f3df3db4a.jpg: 416x416 16 jellyfishs, 8.2ms image 63/63 /content/datasets/fish-1/test/images/IMG_8599_MOV-0_jpg.rf.576be46281797dc6ac8343125cfc1895.jpg: 416x416 8 jellyfishs, 8.2ms Speed: 0.3ms pre-process, 8.5ms inference, 1.0ms NMS per image at shape (1, 3, 416, 416) Results saved to runs/detect/exp
In [ ]:
#display inference on ALL test images
import glob
from IPython.display import Image, display
for imageName in glob.glob('/content/yolov5/runs/detect/exp/*.jpg'): #assuming JPG
display(Image(filename=imageName))
print("\n")
Results
In [ ]:
#Display the graphs for inference
from IPython.display import Image, display
display(Image('/content/yolov5/runs/train/exp/F1_curve.png'))
From the F1-confidence curve, the confidence value that optimizes the precision and recall is 0.418.In many cases higher confidence value is desirable.F1 curve is basically how well our detector is performed.
In [ ]:
display(Image('/content/yolov5/runs/train/exp/PR_curve.png'))
The precision recall curve shows the tradeoff between precision and recall of different threshold.A high area under the curve represents the high recall and high precision,high precision means low false positive rate, high recall relates to a low false negative rate. our curve as close as possible to the top right corner that means our model performed well.
In [ ]:
display(Image('/content/yolov5/runs/train/exp/P_curve.png'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/R_curve.png'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/confusion_matrix.png'))
Confusion matrix is a summary of prediction results on classification problem.The number correct and incorrect predictions are summarized with count values and broken down by each class.The highest prediction is for star fish which is 0.95% and jellyfish 0.83% and penguin 0.81% and so on.
In [ ]:
display(Image('/content/yolov5/runs/train/exp/results.png'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/labels_correlogram.jpg'))
A correlogram or correlation matrix allows to analyse the relationship between each pair of numeric variables of a dataset.A realtionship between each pair is visualized by scatterplot.our correlogram performed well beacuse of high scattering.
In [ ]:
display(Image('/content/yolov5/runs/train/exp/labels.jpg'))
Display the training batch
In [ ]:
display(Image('/content/yolov5/runs/train/exp/train_batch0.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/train_batch1.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/train_batch2.jpg'))
Display the validation batch
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch0_labels.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch0_pred.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch1_labels.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch1_pred.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch2_labels.jpg'))
In [ ]:
display(Image('/content/yolov5/runs/train/exp/val_batch2_pred.jpg'))
Discussion
We sucessfully trained our object detection using YOLOV5 for Aquariums.Aquatic animals play an important role for the environment and humans daily usage.The importance of aquatic animals come from the part of that they provide food,medicine,Energy shelter and raw materials that are used for our daily life.By using this computer vision and Deep learning model we can monitor the sea animals and keep the population under check and protect them from Extinction.
Interpretation
I trained 400 epochos for the best results but my model trained only upto 293 epochos because the best results are observed at epochos 193.I thought best result will be observed between 300 to 400.
Mean average precision is commonly used to analyze the performanace of object detection models. so MAP(mean avearge precision) for the overall model is 0.79% which is pretty good and our model performed very well.
Mean average precision by classes
- Fish 0.72%
- jellyfish 0.86%
- penguin 0.84%
- puffin 0.64
- shark 0.80%
- starfish 0.93%
- stingray 0.77%
The best performed class is starfish which is 0.93% and least performed class is puffin 0.64%.
Literature citation
https://github.com/ultralytics/yolov5
https://app.roboflow.com/kk-fgzul
https://blog.roboflow.com/yolov5-improvements-and-evaluation/
https://blog.roboflow.com/mean-average-precision/#what-is-the-precision-recall-curve
https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data
https://pytorch.org/hub/ultralytics_yolov5/#:~:text=YOLOv5%20%F0%9F%9A%80%20is%20a%20family,Model